TeEFusion: Blending Text Embeddings to Distill Classifier-Free Guidance

TL;DR

TeEFusion is a distillation method that efficiently mimics complex guidance strategies in text-to-image models, enabling faster inference without sacrificing image quality.

Contribution

It introduces a simple fusion of text embeddings to replicate guidance effects, reducing inference costs in text-to-image synthesis.

Findings

Student models achieve up to 6x faster inference

Maintains comparable image quality to teacher models

Effective distillation of complex sampling strategies

Abstract

Recent advances in text-to-image synthesis largely benefit from sophisticated sampling strategies and classifier-free guidance (CFG) to ensure high-quality generation. However, CFG's reliance on two forward passes, especially when combined with intricate sampling algorithms, results in prohibitively high inference costs. To address this, we introduce TeEFusion (Text Embeddings Fusion), a novel and efficient distillation method that directly incorporates the guidance magnitude into the text embeddings and distills the teacher model's complex sampling strategy. By simply fusing conditional and unconditional text embeddings using linear operations, TeEFusion reconstructs the desired guidance without adding extra parameters, simultaneously enabling the student model to learn from the teacher's output produced via its sophisticated sampling approach. Extensive experiments on state-of-the-art models such as SD3 demonstrate that our method allows the student to closely mimic the teacher's performance with a far simpler and more efficient sampling strategy. Consequently, the student model achieves inference speeds up to 6$\times$ faster than the teacher model, while maintaining image quality at levels comparable to those obtained through the teacher's complex sampling approach. The code is publicly available at https://github.com/AIDC-AI/TeEFusion.